Fault Monitoring and Classification Method of Rolling Bearing Based on KICA and LSSVM

Abstract:

The running process of rolling bearing is often nonlinear and abnormal. Therefore, this paper uses the method which combing KICA and LS-SVM to achieve bearings' fault monitoring and classification. Firstly, the vibration signal is mapped into the high dimensional space by using kernel methods, constructing the I2, Ie2 and SPE indicator in the high dimensional space to monitor the process data. And then when the fault occurred, extracting the time domain and wavelet energy features to construct the multi-domain mixed feature set, and input the feature vector into LS-SVM for classification. Experimental results show the method can complete the fault monitoring and classification of rolling bearing effectively.

• Periodical: Advanced Materials Research (Volumes 971-973)
• Pages: 476-480
• DOI: https://doi.org/10.4028/www.scientific.net/AMR.971-973.476
• Online since: June 2014
• Authors: Bao Ma*, Jian De Wu, Jun Ma, Xiao Dong Wang, Yu Gang Fan
• Keywords: Fault Classification, Fault Monitoring, KICA, Ls-SVM, Multi-Domain Mixed Feature

References

[1] Jin Chen: Mechanical vibration monitoring and fault diagnosis of equipment (Shanghai Jiao Tong University Press, Shanghai 1987).

[2] Jong-Min Lee: Statistical process monitoring with independent component analysis. Journal of Process Control. Vol. 14 (2004), p.467–485.

DOI: 10.1016/j.jprocont.2003.09.004

[3] Zhong-gai Zhao, Fei Liu: Nonlinear Process Monitoring Method Based on Kernel Independent Component Analysis. Journal of System Simulation. Vol. 20(2008), pp.5585-5588.

[4] Xi Zhang, Wei-wu Yan, in: Nonlinear On-line Process Monitoring and Fault Detection Based on Kernel ICA, pp.222-227 of ICIA, Information and Automation Publishers(2006).

DOI: 10.1109/icinfa.2006.374116

[5] Kui-he YANG, Gan-lin SHAN, Ling-ling ZHAO: Steam turbine fault diagnosis based on least squares support vector machine. Control and Decision. Vol. 22(2007), pp.778-782.

[6] Tao Peng, Hui-bin Yang, Jian-bao LI: Mixed-domain feature extraction approach to rolling bearings faults based on kernel principle component analysis. Journal of Central South University. Vol. 42(2011), pp.3384-3391.

[7] Ya-guo Lei, Zheng-jia He, Yah-yang Zi: Fault diagnosis based on novel hybrid intelligent model. Chinese Journal of Mechanical Engineering. Vol. 44(2008), p.112—117.

DOI: 10.3901/jme.2008.07.112

[8] Sen Deng, Bo Jing, Wei Zhou: Multi-stage FastICA algorithm with high convergence rate. Chinese Journal of Scientific Instrument. Vol. 32(2011), pp.2430-2436.

[9] Xisen Wen: Pattern recognition and condition monitoring (Science Press, Beijing 2007).

[10] R. RUBINI, U. MENEGHETTI: Application of the envelope and wavelet transform analyses for the diagnosis of incipient faults in ball bearings. Vol. 15(2001), pp.287-302.

DOI: 10.1006/mssp.2000.1330

[11] Tong-hua Ling, Yan-cheng Liao, Sheng Zhang: Application of wavelet packet method in frequency band energy distribution of rock acoustic emission signals under impact loading. Journal of vibration and shock. Vol. 29(2010), pp.127-130.

[12] Information on http/www. eecs. cwru. edu/laboratory/bearing/download. htm.